A Study on Lung Cancer Identification Using Extra Trees-Based Model

• Published in: Journal of computer science and control systems Vol. 15; no. 1; pp. 9 - 13
• Main Authors: Moruff, Oyelakin Akinyemi, Bolaji, Ogidan, I, Olufadi H, Aolat, Raheem-Giwa, H, Buraimah O, M, Rilwan D
• Format: Journal Article
• Language: English
• Published: Oradea University of Oradea 01.05.2022
• Subjects: Accuracy; Algorithms; Artificial intelligence; Classification; Classification schemes; Data analysis; Datasets; Decision trees; Feature selection; Lung cancer; Lung diseases; Machine learning; Medical prognosis; Pathology; Recall; Software; Support vector machines
• ISSN: 1844-6043; 2067-2101

Cancerous diseases come in different forms. One of the most debilitating types of cancerous-related ailments is lung cancer. The disease has been reported to be the second largest type of cancer among men and women in the world. Aside using medical diagnosis and treatment to attend to lung cancer issue, a sub-field of Artificial Intelligence popularly called machine learning has been found to be promising for identification of the cancer. This study investigated the performances of a three-based machine learning model for the identification of lung cancer in public labeled lung dataset. The dataset was collected from UCI machine learning repository and it contains different features on lung related ailments. Prior to building the model, feature preprocessing and selection tasks were carried out. The tree-based learning classifier used is Extra Trees. The algorithm was trained and tested using the pre-processed dataset. An accuracy, precision and recall were used as he metrics for the evaluation. The results obtained for accuracy, precision and recall are: 99.16%, 98.82% and 98.88% respectively. Thus, the experimental results showed that the tree-based lung classification model had promising results when compared to related studies.